Credo's Insiders Are Selling into the Strength. That Is a Clue, Not a Confession.

The most-read document in Credo Technology's filing trail this month was probably the shortest one. It is the Form 144 — the notice an officer files when they intend to sell their own stock — and the headline it produced read like a warning: an officer proposing to sell about $5 million of Credo shares. If you have never read a Form 144, you are reading the right thing, because the document tells a quieter and more useful story than the headline.

First, the scale. That entire proposed sale is less than one-fiftieth of one percent of Credo's roughly $28 billion market value. A single five-million-dollar print would have a hard time moving a stock that trades more than ten million shares on an ordinary day. If you pulled up the filing expecting a boardroom betrayal, the first takeaway is that one sale, by itself, is almost nothing.

So why do these forms keep running through the news feed? Because Credo's own people are filing a lot of them.

The sellers, and what they've actually done

The sales are not theoretical. On July 14, 2026, Credo's chief technology officer and co-founder, Chi Fung Cheng, sold 27,500 shares for roughly $6.6 million, at an average price near $240. It was not a spur-of-the-moment decision to bail: the sale ran through a Rule 10b5-1 trading plan adopted back on September 5, 2025, executed by the Cheng Huang Family Trust. In other words, the decision was locked in nearly a year before the shares changed hands.

The same person filed another sale on September 4, 2026, this time 3,790 shares. The pattern repeats up and down the executive suite. President and CEO William Brennan sold about $10.5 million of stock in December 2025 under a pre-arranged plan, and related filings this year have listed as many as 275,000 shares for future sale through the same kind of 10b5-1 structure. Over the last three months, insiders have bought nothing while selling several million dollars' worth.

None of this is illegal, and none of it is unusual. A 10b5-1 plan is just an agreement an insider signs to sell on a fixed schedule so that the timing cannot be accused of trading on nonpublic information. The reason the forms keep appearing is mechanical: when a company's stock runs up by hundreds of percent in a year, the people who hold the largest stakes get a lot of their pay and their wealth in the form of shares, and they routinely convert a slice of that into cash. Seen that way, a co-founder selling a small percentage of a roughly $1.4 billion personal holding is diversification, not desertion.

That is the benign reading, and it is real. But the evidence ladder does not stop at "benign." When the insiders who know a company best are selling consistently into strength and the price keeps setting records, the honest investor notes the direction of the flow even while acknowledging how ordinary it is.

What the selling does not mean

Before weighing what the sales say about price, it is worth checking what they do not say about the business — because a reader could easily mistake insider selling for evidence the numbers are fake. The cash record clears Credo on that count.

The company just reported its first fiscal quarter of 2027, for the period ended August 1, 2026: revenue of $479 million, up 114.7% year over year. A year before, fiscal 2026 as a whole produced $1.335 billion in revenue, more than triple the prior year, with non-GAAP net income of $662 million. The gross margin came in near 68% on a non-GAAP basis, and the quarter ended with $764 million in cash and short-term investments, with no meaningful debt. The operating cash flow is real, and the balance sheet is a lender's dream.

This is the moment the detective in me usually finds a contradiction between reported profit and actual cash. Here the two line up: the growth is genuine, funded by customers, and corroborated by cash on the books. The selling is not a confession that the accounting is a mirage. If the share price were the whole game, the temptation would be to scream "insiders out, collapse coming," but that is short-seller theater, and this record does not support it.

The lesson is about price, not deception

Which leaves the actual question the Form 144 should force you to confront: not whether Credo is a real winner, but whether the price you would pay today still has room to fall even if the company keeps winning.

Consider where the shares sit. Credo peaked near $309 over the past year and has since slid toward $150 — down around 50% from that high even as the company delivered a 115% growth quarter. At the current price, the stock trades at roughly 18 times trailing revenue and about 52 times trailing earnings. That is the market paying for perfection: for the numbers to justify the multiple, the company cannot merely keep growing; it has to keep growing at a pace that outruns one of the highest prices in the sector. Fellow AI-networking names command high multiples too, but Credo's sits at the very top of the range.

Here is the uncomfortable symmetry. The co-founder sold stock at $240 in July. In September the same stock changed hands in the $150s. The insiders who understand their business better than anyone on a message board chose to harvest hundreds of millions of value all the way up — and, until now, have chosen not to buy a single share on the way down. Insiders sell for any number of reasons, but they buy for only one: they think the stock is cheap. Not one of Credo's leaders has apparently made that call recently.

That is not a forecast that the company is doomed. It is a warning about whose edge you are trading against. When you buy a stock that insiders are systematically selling and that trades at 52 times earnings, you are not buying on the same information terms as the people inside the building. You are paying a full price for growth that must compound nearly flawlessly to earn it back.

Three ways this resolves, and what you pay

Set aside the headline and run the scenarios.

Benign and continuing. The sales are pre-planned, diversified, and tiny against the float; the company compounds, and the multiple slowly normalizes as the business matures into a larger revenue base. Sellers' diversification cost the buyer nothing beyond the ordinary ups and downs. This is the outcome the filing itself most resembles.

Persistent lawful pressure. Credo's growth is real but depends on a handful of hyperscale data-center buyers and on an AI capital-spending cycle that will eventually face harder year-ago comparisons. If growth fades from triple-digit to merely strong, a 52-times-earnings stock can fall hard even while the business stays excellent — and the insiders, who saw the easy re-rating, keep trimming. This is the case where a holding period can disappoint despite good operations.

Materially misstated. For this to be true, the cash and the customers would have to be fiction, and nothing in the filed statements supports that. The honest statement of this case is that it has no evidentiary support. The real tail risk here was never fraud; it is concentration in a few big customers and the timing of the AI buildout.

The shareholder invoice, then, is not a hidden accounting charge. It is the price of the multiple itself. Buying Credo today at $150 — down half from the high, yet still 52 times earnings — means you are assuming the sellers are wrong to sell. The strongest available evidence points the other way: the numbers are real, the growth is real, and the people who know the business best are treating this price as one worth cashing in, not adding to.

The filing that will actually move this case one rung on the evidence ladder is the next quarter's release in early December, and the quarters after it. Watch not the headlines about five-million-dollar sales, but whether revenue keeps outrunning that 52-times multiple — and whether, at some point, a single insider finally decides the stock is cheap enough to buy.